sign in sign up
<<
Home , Bioluminescence, Dominance (ecology), Nekton, Phytoplankton, Productivity (ecology), Pyrocystis fusiformis

Bioluminescence to reveal structure and interaction of coastal planktonic communities

Mark A. Moline a , Shelley M. Blackwell a, James F. Case b, Steven H.D. Haddock c, Christen M. Herren c, Cristina M. Orrico b, Eric Terrill d a Biological Sciences Department, California Polytechnic State University, 1 Grand Avenue, San Luis Obispo, CA 93407, USA b Marine Science Institute, University of California Santa Barbara, Santa Barbara, CA 93106, USA c Monterey Bay Research Institute, Moss Landing, CA 95039, USA d Marine Physics , Scripps Institution of Oceanography, La Jolla, CA 92093, USA

abstract

Ecosystem function will in large part be determined by functional groups present in biological communities. The simplest distinction with respect to functional groups of an is the differentiation between primary and secondary producers. A challenge thus far has been to examine these groups simultaneously with sufficient temporal and spatial resolution for observations to be relevant to the scales of change in coastal . This study takes advantage of general differences in the flash kinetics between planktonic dinoflagellates and to measure relative abundances of the two groups within the same-time space volume. This novel approach for distinguishing these general classifications using a single sensor is validated using fluorescence data and exclusion experiments. The approach is then applied to data collected from an autonomous underwater vehicle surveying 4500 km in Monterey Bay and San Luis Obispo Bay, CA during the summers of 2002–2004. The approach also reveals that identifying trophic interaction between the two planktonic communities may also be possible.

1. Introduction

Coastal regions are responsible for approximately 30% of global cycling thus occur when these highly concentrated predator and (Holligan and Reiners, 1992) and, as such, are prey fields intersect. Because of the high rates and levels of zones of the highest biogeochemical cycling per area with respect activity in coastal systems, the mechanisms governing patch to , , phosphorus, and trace metals (Ducklow and distribution and coherence of and their biological McCallister, 2005; Jahnke, 2005). Mediation, persistence, and interactions are major topics of ongoing research. variability of these rates of biogeochemical cycling are primarily While clearly important, assessing the distribution of driven by the structure and activity of biological communities. and particularly their interactions have been challenges for These communities are, in turn, organized non-randomly, and can oceanographers. This, in part, stems from the array of approaches be layered relative to the physical structure of and used to quantify planktonic communities in situ, and the distribution of , by advective processes and by behavioral distinction between approaches specific for zooplankton versus differences within and between organisms (Deutschman et al., . Bio-optical approaches, such as fluorometry, have 1993). Because of these varied mechanisms for accumulation (or successfully delineated autotrophic populations and communities patch formation) of different planktonic organisms, their hor- in situ for many years (Yentsch and Menzel, 1963; Lorenzen, 1966). izontal and vertical distributions are often heterogeneous, vary More recently, in situ absorption has been used as a tool to assess between organisms, and are scaled to the physical, chemical, and phytoplankton concentrations (Moore, 1994), as well as separate biological forcing. The size of these patches also generally scales out specific functional groups (Schofield et al., 2004) or , inversely to the organisms’ size (Levin, 1992), with the largest such as harmful (Kirkpatrick et al., 2000), based on their patches represented by autotrophic phytoplankton, and less pigment signatures. Similarly, ocean acoustic approaches have concentrated larger heterotrophic organisms in successively been developed to map zooplankton and (Johnson, 1948; smaller patches (Hall and Raffaelli, 1993). Maximal trophic Holliday and Pieper, 1980; Flagg and Smith, 1989). Recent interactions, transfer of carbon, and rates of biogeochemical advancements in multi-frequency acoustics have been able to distinguish between zooplankton groups and species (Holliday Another body of literature has attempted to examine spatial et al., 1989; Pieper et al., 1990; Cochrane et al., 2000; Benoit-Bird and temporal variability in bioluminescence from a and Au, 2003). While both bio-optical and acoustic approaches and ecosystem perspective. In approaching this problem, in situ provide significant information on the distribution and concen­ sensors, called bathyphotometers, are employed to quantify the tration of phytoplankton and zooplankton at a range of sizes and amount of BP and community structure of bioluminescent groups, in situ measurements are rarely concurrent and at organisms in a particular body of water and relate these patterns different scales, making synthesis and integration difficult. to the local- or ecosystem-level dynamics. Case et al. (1993) and Additionally, uncertainty in the spatial and temporal intersection Alberte (1993) review the development of bathyphotometers and of these communities limits the extent of our understanding of patterns of oceanic bioluminescence. Although developed in the trophic interaction and thus the , rates, and scales of 1950s, one of the first large-scale applications of bathyphot­ coastal biogeochemical cycling. These uncertainties have also ometers took place in late 1980s in North Atlantic to examine the played a role in limiting the extent to which has been differences in production by various planktonic taxa integrated into dynamical regional ocean models, which have (Batchelder and Swift, 1989; Losee et al., 1989; Batchelder et al., significantly advanced with respect to physical oceanography 1990, 1992; Swift et al., 1995). These measurements are becoming (Kantha and Clayson, 2000). A measurement is therefore needed more prevalent and have now been conducted off , on that can provide simultaneous data for different planktonic profiling and undulating systems, on moorings, and on autono­ communities at time and space resolutions similar to routine mous underwater vehicles (AUVs; Widder et al., 1993; Moline oceanographic parameters, such as temperature, salinity, and et al., 2001, 2005; Herren et al., 2005). These efforts have provided fluorometry. new insight into the distribution of coastal bioluminescence at The measure of bioluminescence or bioluminescence potential ecosystem scales as it relates to physical forcing and physiological (BP) has been reported in the literature for some time (Clarke and rhythms (Widder et al., 1999; Shulman et al., 2003, 2005; Moline Wertheim, 1956; Clarke and Kelley, 1965; Seliger et al., 1969). et al., 2005). Early research on this phenomenon was driven primarily by the From this body of work, there have been a number of general desire to understand physiological mechanisms for biolumines­ relationships that have been derived from the measurement of cence, as well as the ecological advantage that bioluminescence marine bioluminescence. For a given planktonic community affords to organisms (Alberte, 1993). Previous work in marine (highly dependent on locale and season), the number of bioluminescence can be divided into a number of categories and bioluminescent organisms and total bioluminescence are propor­ depends largely on the level of organization. Bioluminescence is tional to the total (Lapota, 1998). The intensity of produced by over 700 genera representing 16 phyla, spanning the bioluminescent flash and duration of flash have also been shown range of small single-cell to large (, to correlate with the size of (Lapota and Losee, 1984; 1987). As such there have been a number of studies examining the Lapota et al., 1992). Because of this general difference, biolumi­ phenomena on individual, population, and ecosystem levels. On nescence flash kinetics can be used to delineate these groups the organism level, the physiological and cellular basis for (Fig. 1). Even though larger bioluminescent organisms generally bioluminescence (Rees et al., 1998), the spectral quality and flash produce more light, in locally stable environments, their numbers kinetics of bioluminescence (Latz et al., 1988; Haddock and Case, are proportionally lower relative to smaller single-celled dino­ 1999), and how these relate to aspects such as circadian rhythms flagellates. The majority of coastal BP scales inversely with the (Soli, 1966; Morse et al., 1989), (Johnson et al., size spectrum, with dinoflagellates generally responsible for the 1998), and diet (Haddock et al., 2001) have been well documented. majority of the signal (70–90%; Lapota et al., 1988; Swift et al., On a population level, bioluminescence has been studied as it 1995). relates to predator avoidance, prey attraction, and intra-species Here we use the general relationship between biolumines­ communication (Burkenroad, 1943; Morin, 1983; Morin and cence flash intensity with organism size to interpret signals Cohen, 1991; Abrahams and Townsend, 1993). measured from a bathyphotometer deployed on an AUV during

1.00E+12

1.00E+11

-1 1.00E+10

1.00E+09 Photons flash

1.00E+08

1.00E+07 0 1000 2000 3000 Size (μm)

Fig. 1. Relationship between bioluminescence flash intensity (photons flash�1) and organism size in dinoflagellates (open symbols) and zooplankton (closed symbols). Data for flash intensity compiled from Lapota and Losee (1984; circles), Lapota et al. (1988; squares), and Lapota et al. (1992; triangles). Organism size ranges estimated from Thomas (1997) and Johnson and Allen (2005). 2002 and the 2003 Autonomous Sampling Observation Network II Herren et al., 2005). Here, a light-baffled photomultiplier tube (AOSN-II) experiment in Monterey Bay and develop a means to (PMT) measures stimulated light between 300 and 650 nm distinguish and delineate the general structure of coastal produced by the entrained organisms. The inside of the chamber planktonic communities and their interactions. This approach is coated with a 0.075-mm flat white coating to maximize the may complement traditional measurements and serve to validate amount of stimulated light measured by the PMT. The PMT was or access uncertainties over relevant scales. configured to take measurements at 2 Hz. The flow rate through the chamber is dependent on the rotation rate of the impeller rotor. This rate is adjusted to achieve residence times of 1.2–1.4 s, � 2. Methods or flow rates of approximately 400 ml s 1. A flow meter monitors pumping rates using a magnet and a Hall-effect sensor to generate 2.1. Bioluminescence measurement a period signal, which is converted to an analog signal of flow rate. The flow rates are measured as water passes from the detection The bioluminescence bathyphotometer used in this study to chamber to exhaust outlets. The bioluminescence bathyphot­ quantify BP is described in Herren et al. (2005). A centrifugal-type ometer was integrated into the front section of a Remote impeller pump drives water into an enclosed 500-ml chamber and UnitS (REMUS) AUV system (Fig. 2A; creates turbulent flow, which mechanically stimulates biolumi­ Moline et al., 2001, 2005; Blackwell, 2002; Herren et al., 2005). In nescence. The measure of BP is therefore an index of the total order to prevent premature stimulation of bioluminescence by the luminescent capacity of organisms in a set water volume. The moving vehicle, water is taken directly through the front nose measure assumes similar flow-stimulation characteristics for section of the vehicle. Two light-baffling turns in the nose serve to different groups of organisms and is dependent on the character­ minimize ambient light contamination. No significant ram effect istics of the bathyphotometer, which can vary significantly (cf. on light production or flow rate from the vehicle itself was found

Fig. 2. (A) The REMUS autonomous underwater vehicle used in this study being deployed in San Luis Obispo Bay, CA. Integrated into the nose section of the vehicle is the bathyphotometer for quantification of bioluminescence (see text). (B) Two bathyphotometers attached to a Schindler trap for method validation in this study. Screened bottles were placed on the exhausts of the bathyphotometers to collect organisms for quantification and identification. (C) The REMUS with similar bottles attached to the bathyphotometer exhaust ports to capture organisms for validation tests. with this integrated system. Two additional bioluminescence while profile sampling occurred off the California Polytechnic bathyphotometers were used in profiling mode as part of State University’s Center of Coastal Marine Sciences pier. All validation tests (see below). Cross-calibration between the three sampling for this study was conducted between 22:00 and 04:00 instruments was ensured using a standard isotropic light source local time as BP is a diurnally dependent measure, but it has been probe inserted into the individual stimulation chambers (Herren shown to be generally stable during this 6-h period (Moline et al., et al., 2005). 2001). In Monterey Bay, the REMUS was programmed to undulate between 4- and 40-m depth at a speed of approximately 2 m s�1. 2.2. Sampling approach Navigation of the AUV was by an internal compass corrected for by onboard-measured 3-D currents (Moline et al., 2005). Naviga­ Data for this study were collected from Monterey Bay in tional error over the combined REMUS runs for this study was August 2002 and 2003 as part of the AOSN-II experiment, and in �1.71 of distance covered. For sampling in San Luis Obispo Bay, the San Luis Bay in June/September 2004 (Fig. 3). Sampling with the REMUS was programmed to travel across a 400-m transect at REMUS in Monterey Bay in 2002 occurred along transect ‘‘a’’ while constant depths of 2 and 6 m along the 12-m isobath. Twenty-mm in 2003, sampling was conducted along transects ‘‘a’’ and ‘‘b’’, nets were attached to the exhaust ports of the REMUS during each approximately 21 km in distance. REMUS sampling was these deployments to capture the organisms going through the conducted in San Luis Obispo Bay along a cross-shore transect, sampling chamber (Fig. 2C). The REMUS was deployed with this

Monterey Bay

37.000° 37.917°

San Francisco a b 36.833° 37.333°

36.667°

36.750° California 36.500°

36.167° ° ° ° -122.333 -122.167 -122.000 -121.833°

San Luis Obispo Bay 35.417° 35.583° Pacific Ocean 35.333°

35.250° 35.000° 35.167°

35.083°

34.417° 35.000°

34.917°

-122.583° -122.000° -121.417° -120.833° ° ° ° ° ° ° °

-121.167-121.083-121.000-120.917-120.833-120.750-120.667

Fig. 3. Study areas (insets) in relation to the central coast of California. Monterey Bay was the site of the REMUS deployments along two transect lines (‘‘a’’ and ‘‘b’’) during August 2002 and 2003. San Luis Obispo Bay was the location for validation tests with the autonomous profiler (circle) and the REMUS in 2004 (black line). net configuration twice in San Luis Obispo Bay along the two biological fields (Fig. 4). The cross-shore transect was character­ depths. During the second deployment, for the purpose of ized by a stratified with evidence of upwelled water excluding larger plankton from the excitation chamber, an out to 5 km. This pattern is consistent with a recurring cyclonic additional 190-mm net was placed between the light-baffling that forms in Monterey Bay during periods nose and the bathyphotometer, 0.25 cm from the impellor to (Shulman et al., 2003). Phytoplankton were layered inshore of prevent pre-stimulation of bioluminescence. Given the flow rate 5 km with a deeper and more diffuse distribution between 5 and of the bathyphotometer (�400 ml s�1) and the diameter of the 10 km offshore. At 10 km, there was a twofold decrease in intake (3.2 cm), the time between the screen and the impellor phytoplankton biomass in surface associated with a (where the stimulation is designed to occur) was 5 ms, which is salinity front. The depth distribution of high BP was also shallower significantly lesser than the flash response latency of these inshore of 5 km and deeper between 5 and 10 km offshore, similar organisms (Widder and Case, 1981). to fluorescence, although the higher BP values were more In San Luis Obispo Bay, a bioluminescence bathyphotometer concentrated at depth. There were areas of high BP measured was attached to a Schindler sampling trap (Fig. 2B). The sampler offshore of 15 km, near the bottom inshore, and on the offshore was suspended at the depth of peak bioluminescence (2.5 m) for side of the front at 5 km that did not co-occur with high values of 3 min. The intake of the bathyphotometer was alternately pre- fluorescence. screened with a 190-mm screen or not screened (see above). In regions where both BP and fluorescence are high, the Initially, a large-mesh �2500-mm pre-screen was also used to traditional interpretation is that the majority of bioluminescent examine effects of pre-stimulation and impact of screening on the community is autotrophic (Lieberman et al., 1987; Lapota, 1998; organism. There were no significant differences in number or type Geistdoerfer and Cussatlegras, 2001). Some late-stage phyto­ of organisms or bioluminescence signal between the large-mesh plankton blooms may yield a successional accumulation of control and the non-screened treatment (data not shown). We autotrophic dinoflagellates that may include bioluminescent therefore report only the 190-mm pre-screen and non-screen species, e.g., Lingulodinium sp. (formerly Gonyaulax sp.) or conditions. In both of these conditions, the exhaust water from Ceratium fusus, which may also contribute to a positive relation­ the bathyphotometer was screened through a 20-mm screen to between a and bioluminescence (Swift et al., capture organisms that traveled through the instrument for 1995; Lapota, 1998). In addition, it has been found that identification and enumeration. There was no visible impact of dinoflagellate blooms increased the amounts of luminescent the bathyphotometer on the structure of either phytoplankton or (Alldredge et al., 1998; Haddock, 1998), which can zooplankton. Plankton were identified in a 100-ml settling be a dominant source of bioluminescence (Herren et al., 2003). chamber using an inverted microscope. Likewise, when bioluminescence is high with little fluorescence, traditional interpretations would suggest a of hetero­ trophic organisms. These assumptions have been shown to be 2.3. Signal processing generally valid, given the large differences across coastal ecosys­ tems in the percent of both and that are The variance and mean BP for a given location were calculated bioluminescent (Lapota, 1998). While this is a common approach from a sliding data window. The size of the sliding window was for delineating plankton communities, it is difficult to apply equivalent to 25 m linear distance traveled by the REMUS vehicle objectively across space and time. As the flash kinetics (intensity for data collected in Monterey Bay and San Luis Obispo Bay and and duration) differ with size, and size generally delineates was objectively determined by identifying the length scales of between phytoplankton and secondary producers (Fig. 1), we variability, detailed in Moline et al. (2005) and Blackwell et al. attempted to use the bioluminescence signal intensity as a single (2007). For the time series tests performed with the profiling measure to identify the coarse structure of the planktonic bathyphotometers a data window of 12.5 s was used (25 community. Dinoflagellates generally have a lower flash intensity observations). The square root of variance of BP and mean BP than zooplankton and, integrated over a large region, are generally were used to generate the coefficient of variation (CV) used in this more abundant in number and uniform in their distribution. study. This approach highlights the differences in bioluminescent Zooplankton (i.e. ), conversely, are fewer in number in an flash intensity rather that flash duration as a means to separate equivalent volume, but have a more intense flash. These dinoflagellates and zooplankton. While flash duration is certainly differences are hypothesized here to lead to variation in signal important and contains species-level information (Widder et al., outputs from the bathyphotometer. 1993), it is problematic for many studies using bathyphotometers as residence times of these instruments vary (cf. Herren et al., 2005) and are shorter than flash durations of many organisms. 3.2. Planktonic communities The simple volume replacement time calculated for the bath­ yphotometer is on the order of a second; however, it is clear that a Using the data collected by the REMUS in Monterey Bay in decreasing number of organisms can be retained within the flow 2002, we compared the average bioluminescence intensity to the field for longer periods (6–10 s; Herren et al., 2005). Because of square root of variance in signals, or CV, as a way to distinguish this uncertainty in retention and the strong correlation these communities. There was, in fact, a distinct bifurcation of the 2 ¼ (R 0.81; exponential fit) between flash intensity and duration data, with one distribution of points showing a higher CV (slope) (Lapota and Losee, 1984), we have focused on intensity in this than the other grouping of points (Fig. 5). Given the assumptions study. above, the grouping with high CV was consistent with zooplank­ ton; decreased flash frequency being however more intense. To attempt to validate this hypothesis, the concurrent fluorescence 3. Results and discussion measures were overlaid on the distribution of bioluminescence data (Fig. 5). was grouped with the lower CV signal, 3.1. Coastal dynamics suggesting these were either autotrophic dinoflagellates or heterotrophic dinoflagellates associated with autotrophic species. Data from the REMUS deployments in Monterey Bay in 2002 High fluorescence was absent in the high-CV data cluster, showed significant coefficient of variability in both physical and suggesting zooplankton. Lapota et al. (1989) demonstrated that 15 C)

-10 ° 14 -20 13 12 Depth (m) -30 11 -40 ( Temperature

33.6 -10

-20 33.55 Salinity

Depth (m) -30 33.5 -40

4 -10 3 -20 2 (RFU)

Depth (m) -30 1 Fluorescence

-40

x 1010

10 ) -10 -1 8 L -1 -20 6 4

Depth (m) -30

2 (photons s Bioluminescence -40 0 5 10 15 20 Distance (km)

Fig. 4. Depth distributions of (A) temperature (1C), (B) salinity, (C) fluorescence (RFU), and (D) bioluminescence (photons s�1 L�1) collected by the REMUS vehicle in Monterey Bay along transect line ‘‘a’’ (Fig. 3) on August 20, 2002. Flight path of the REMUS vehicle is shown in (A).

x 1010 9

8 4

7 3.5

6 3

5 2.5

ariance of Bioluminescence 4 2

3 1.5 Fluorescence (RFU)

2 1

1 0.5 Square Root of the V

0 02 4 6 810 12 14 Average Bioluminescence (photons s-1 L-1) 1010

Fig. 5. Average bioluminescence (photons s�1 L�1) as a function of the square root of variance of bioluminescence for data shown in Fig. 4. Variance and average were calculated using a sliding data window of observations, representing 25 m linear distance. Each observation of bioluminescence is overlaid with the concurrent value of fluorescence (RFU). Table 1 Table 2 Identification and (number L�1) of phytoplankton and zooplankton and Identification and abundance (number L�1) of zooplankton and dinoflagellates in in samples collected by Schindler trap in Monterey Bay, CA on August 20, 2002 samples collected through the bathyphotometers in San Luis Obispo Bay, CA on along transect line shown in Fig. 4 June 9, 2004 between 23:45 and 00:31 PDT

Distance offshore (km) 0.5 1.0 1.0 6.7 11.9 11.9 16.1 16.1 21.7 21.7 No screen 190 mm prescreen Depth (m) 6 9 20 9 5 37 7 35 8 35 Zooplankton Dinoflagellates Copepods Autotrophic Metridia sp. a 29 (18) – G. sanguineum 530 240 Calanoid 31 (11) – sp.a,b 10 10 10 30 350 540 120 120 Cyclopoid 4 (2) 5 Ceratium fususb 20 20 20 Nauplii 68 (27) 33 (9) Ceratium sp. b 20 20 20 10 10 150 100 Alexandrium cantenellab 20 20 20 120 Other a Prorocentrum micans 20 20 20 50 50 310 230 Siphonophore 2 (1) – Prorocentrum sp. b 50 30 30 20 20 420 320 18 (5) 3 (3) Gyrodinium sp.a 10 10 10 10 420 10 Polychaete larva 30 (16) 22 (7) Pyrocystis sp.b 130 20 larva – 3 L. polyedrab 220 270 Dinoflagellates Oxytoxum sp. 40 Alexandrium sp. a 600 (500) 300 (300) a Heterotrophic Ceratium furca 100 (100) 0 a Protoperidinium sp. b 130 160 20 sp. 100 (100) 200 (100) a Polykrikos schwartzii 10 Gonyaulax sp. 600 (200) 700 (400) a Noctiluca scintillansb 20 20 Gymnodinium sp. 400 (300) 0 a Oxyphysis sp. 30 Lingulodinium polyedrum 0 100 (100) a Dinophysis sp. b 10 30 Protoperidinium 5800(1500) 5100(1800) Other 400 (0) 0 sp. 10 10 10 10 10 10 10 10 Samples were collected from a bathyphotometer unscreened and prescreened with Eucampia sp. 10 10 10 100 10 100 100 10 190 mm mesh. The numbers are the totals of three replicate samples. Numbers in Pseudonitzschia sp. 10 10 100 10 100 1000 100 1000 1000 100 parenthesis are the subset from one of the trials shown in Fig. 6. Thalassionema sp. 10 10 10 10 10 10 10 10 a Bioluminescent or can have bioluminescent species. Coscinodiscus sp. 10 10 10 10 10 10 10 10 10 10 Other 70 30 10 30 30 50 20 70 40 80 10 Zooplankton 80 30 30 160 110 110 470 80 70 70 9 10 10 10 Copepod Nauplii 10 20 20 20 20 60 40 20 20 8 Veliger 10 10 Pluterus larvae 10 7

a Species can be heterotrophic or mixotrophic. 6 b Bioluminescent or can have bioluminescent species. 5

even when correlations between fluorescence and biolumines­ 4 cence are strong, it does not necessarily confirm that the 3 fluorescence has been due to the luminescent organisms as heterotrophic dinoflagellates may often dominate the planktonic Percent of Observations 2 community. The plankton collected along the transect clearly showed the majority of phytoplankton cells were diatoms; 1 however, a high fraction of the cells were dinoflagellates and a 0 significant portion of those were bioluminescent (Table 1). Of the 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 � bioluminescent fraction, 70% were autotrophic. It is clear that all Coefficient of Variation the fluorescence is not related to luminescent organisms with lower CV; however, the amount of fluorescence in this group is Fig. 6. Coefficient of variation (CV) for bioluminescence made in San Luis Obispo higher than seen in the high-CV distribution. Therefore, for this Bay, CA on June 9, 2004 between 23:45 and 00:31 PDT. Variance and average location and time, fluorescence provides some confirmation that bioluminescence were calculated using a sliding window of 25 measurements or 12.5 s. A bathyphotometer was held at 2.5 m in the bioluminescence maxima for the variance in bioluminescence measurements can discriminate 3 min. Black circles represent data collected with a 190-mm screen in front of between planktonic communities in this data set. intake; white circles had no net covering intake. The ratio is significantly less with To further validate the use of CV, a number of controlled 190-mm screening covering intake (t-, po0.00, n ¼ 599). experiments were conducted. In June 2004, a bathyphotometer was suspended in the water column in San Luis Obispo Bay (Figs. 2 approach spatially (Fig. 2C; see Methods). The vehicle went on and 3). The bathyphotometer was alternately pre-screened with a two identical missions within an hour of each other; one mission 190-mm screen to exclude zooplankton or not screened. Micro­ with 190-mm screen at the water intake and the other without. scopic identification of the samples going through the bath­ The vehicle maintained two depths over the mission to sample yphotometer confirmed this approach (Table 2). The CV of above and below the (Fig. 7A). The bioluminescence screened bathyphotometer measurements was lower and sig­ was almost twofold higher at depth and the difference in BP nificantly different than the non-screened condition containing measured with (Fig. 7B) and without the screen (Fig. 7C) over the zooplankton community (Fig. 6). This experiment was repeated in intake indicated that signals at depth were generated by triplicate with the same findings. A similar approach was used organisms larger than 190 mm. The CV of bioluminescence signal with the REMUS vehicle in September 2004 to validate the also indicated the presence of zooplankton at depth with a 0

14.5 C) ° 14 5 13.5 Depth (m) 13 emperature ( T 10 x 1010 0 4 ) -1 3 L -1 5 2 Depth (m) 1 (photons s Bioluminescence 10 x 1010 0

4 ) -1 L

3 -1

5 2

Depth (m) 1 (photons s Bioluminescence 10 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Distance (km)

Fig. 7. Results of REMUS deployment in San Luis Obispo Bay in September 2004. The linear distance represents the roundtrip mission of the AUV along the transect. As seen in the data traces, the vehicle traveled to the beginning of the transect about �280 m from the start position, dove, and maintained the vehicle at 2 m to the end of the transect. On the return, the vehicle dove and maintained operation at 6 m until it returned to the beginning of the transect (�600 m), then returned to the start position. This was repeated twice with both traces of temperature shown in (A). Bioluminescence from each deployment is shown separately without prescreening (B) and with a 190-mm prescreen (C). The differences in the CV between 2 m (0.1670.04) and 6 m (0.2370.11) highlighted in panel (B) were significantly different (t-test; po0.01; n ¼ 763). significant difference between depths (Fig. 7). Given the validation progressed, the depth and offshore extent of fluorescence of this technique with concurrent measures of fluorescence, and increased, which is consistent with the physical dynamics. temporal and spatial exclusion experiments with screens, we Intermittent high fluorescence was evident in the center of the applied the approach to the larger AOSN-II data set. bay extending to 30 m, and corresponded to the deepening of the thermocline. Bioluminescence distributions and dynamics showed simila­ 3.3. Dynamic structure of planktonic communities rities with fluorescence, with peak values of 2.3 � 1010 photons s� 1 L�1 along the shoreline (Fig. 8B). The temporal pattern of Data from nine successive nighttime transects across Monterey entrainment into the upwelling front was also evident along Bay show spatial differences with depth and distance offshore, as transect ‘‘a’’, with the BP signal deepening and extending from 5 to well as time of physical and biological structure of the 10 km offshore. The bioluminescent communities in the northeast bay (Fig. 8A). The atmospheric forcing and physical dynamics in appeared concentrated along the coast during this process when this region are well characterized for this time period. The time compared to the initial condition, where the communities sequence of data collected by the REMUS catches a slow transition extended �10 km offshore. There was high BP from 20 to 40 m from a strong upwelling event to a relaxation condition (Shulman in the center of the bay, extending inshore at the beginning of the et al., 2005). Upwelling, affecting the study area, occurred along experiment with the highest signal below the thermocline and the coast north of the bay. Upwelled water entered the southern fluorescence layer. While there were high BP signals in the center part of the bay and displaced coastal water from north and of the bay during the entire study, their distribution and intensity northeastern sections of the bay, where the sampling took place changed significantly. Most evident was the apparent separation (Fig. 3). This effectively set up a cyclonic eddy that pushed water between the nearshore surface BP signals and the deeper signals onto the coast. This was clear in the temperature data beginning in the middle of the bay as the upwelling intensified. This was on the fourth night of sampling (August 13, 2004), where the perhaps due to the intensification of the eddy, as suggested by the thermocline on both shorelines shallows (Fig. 8A). As time vertical distribution of BP, and to some degree fluorescence, on the progressed in the sampling, the thermocline deepened on the last two sampling days (Figs. 8A and B) and the strength of the shore side of transect ‘‘a’’, while remaining relatively shallow on eddy (Shulman et al., 2005). The oscillations in depth distribu­ transect ‘‘b’’, consistent with the entrainment of bay water along tions of the physical and biological parameters, for example the upwelling front to the west of transect ‘‘a’’. Fluorescence data inshore on transect ‘‘a’’ on the last night, are consistent with showed higher concentrations along the coastline, with significant internal waves known to persist in this area (Petruncio et al., layering of the phytoplankton community (Fig. 8A). The offshore 1998). extent of fluorescence distribution was similar to the data The depth distribution of CV of the bioluminescence signal collected in 2002, where higher values were generally restricted during the AOSN-II experiment is consistent with the dynamics to the inner 5 km along the shelf break (Fig. 4). As time described above, with low CV indicative of dinoflagellates, Temperature (°C) Fluorescence (RFU) 10 12.5 15 0 1 2 3

-10 -20 -30 Depth (m) August 10 -40 August 11 August 12 August 13 August 14 August 15 August 16 August 17 August 18

0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40 Distance (km) Distance (km)

Fig. 8A. Depth distributions of temperature and fluorescence collected by the REMUS AUV along the ‘‘a’’ and ‘‘b’’ transect lines in Monterey Bay (see Fig. 3) on nine successive nights from August 10 to 18, 2003. Distance is total distance traveled starting on the north end of transect ‘‘a’’ and finishing on the northeast end of transect ‘‘b’’. The dotted line separates the two transects, indicating the furthest point offshore and the deepest point over Monterey Canyon. The fluorometer on board the AUV had a poor connection with the vehicle during the night of August 14, 2006. It is included as some of the coastal features are still evident. restricted primarily to the coast and corresponding to regions of the bay. While CV remained high under the nearshore fluores­ high fluorescence. The highest CV values were either below high cence, it decreased in the center of the bay as the distribution of fluorescent areas nearshore at the beginning of the experiment or bioluminescent organisms became more uniformly distributed distributed throughout the center of the bay and the thermocline both vertically and horizontally, decreasing the variance in the interface. What was also clear in CV distribution was the gradual signal. Whether the separation between nearshore and bay separation of nearshore communities from those in the center of communities was simply displacement driven primarily by the -1 1 Bioluminescence (photons s L- ) Coefficient of Variation 1e10 3e10 0 3

-10 -20 -30 August 10 Depth (m) -40 August 11 August 12 August 13 August 14 August 15 August 16 August 17 August 18 0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40 Distance (km) Distance (km)

Fig. 8B. Same as 8A except with bioluminescence and coefficient of variation (CV). CV was not calculated where the average bioluminescence was less than 5 �109 photons s� 1 L�1 (black).

circulation pattern or by behavior, it is clear that the potential for 3.4. Applications coupling and trophic interaction decreased as the upwelling intensified. Using this approach for separating dinoflagellates and There are several scenarios where the application of CV and its zooplankton, the total BP over the 9 days was proportioned as 66% interpretation could be difficult. First, as CV is a ratio, the average and 34%, respectively, similar to previous findings (Lapota et al., bioluminescence signal needs to be sufficiently above the back­ 1988; Swift et al., 1995). ground measured by the instrument as evident in Fig. 8B. Second, 10 -5

-10 ) 3 -1

-15 L -1 -20 5 2 -25 Depth (m) -30 (photons s

Bioluminescence 1

-35 Fluorescence (RFU) 1 -40 0

-5 -10 2 3

-15 riation -20 2 -25 Depth (m) -30 ficient of Va 1

-35 Fluorescence (RFU) -40 0 Coef 0 0 5 10 15 20 0 5 10 15 20 Distance (km) Distance (km)

Fig. 9. Depth distributions of (A) bioluminescence (photons s�1 L�1), (B) fluorescence (RFU), (C) the coefficient of variation (CV), and (D) fluorescence (RFU) where CV was greater than 0.6 collected by the AUV along the ‘‘a’’ transect in Monterey Bay on August 20, 2002.

although not apparent in these data, if the communities are boundary conditions. This is known for the physical , but thoroughly mixed the ratio will not be able to differentiate is also true, and most likely different, for the biological commu­ between groups. In addition, some dinoflagellate species can be nity structure in a regional context. mixotrophic or heterotrophic, but exhibit the same spatial distributions and flash kinetics as autotrophic dinoflagellates (Lapota et al., 1988), making the application of CV as an absolute 3.5. Planktonic interaction method of delineating trophic status of a community problematic. Third, the percent of a given phytoplankton or zooplankton In addition to discriminate between planktonic communities, community that is bioluminescent has been shown to vary data collected in 2002 suggest that this approach has potential to significantly in time and space (Lapota, 1998), making CV as a address the interaction of the two groups. As evident in Fig. 5, quantitative measure that is universally applicable improbable. there was a distinct low but slightly elevated fluorescence signal Lastly, like the measure of apparent optical properties (i.e. in the high-CV cluster attributed to zooplankton. Two probable irradiance) restricting sampling during daylight hours, biolumi­ conditions could account for this data distribution. The first nescence requires sampling at night. As shown here, despite these possibility is that the zooplankton were mixed with low real limitations, this single measure in different coastal regions phytoplankton biomass. However, this appears unlikely because and at different times of the year may provide qualitative and in the signal was uniform with no elevated fluorescence values in the some cases quantitative separation between dinoflagellates and high-CV data grouping. Additionally, there was a clear separation zooplankton. in fluorescence, with little to no fluorescence signal between the Rapid delineation of these groups in the field could serve to two CV distributions (Fig. 5). The second possible explanation is significantly advance the integration of biology into dynamic that the fluorometer on board the REMUS AUV was detecting ocean models. Data-assimilative hindcast/forecast and nowcast fluorescence from the zooplankton guts. Zooplankton gut fluor­ models are beginning to couple simple biological models that escence has been a standard measurement for quantifying depend on time and space knowledge of growth and loss terms ingestion, grazing, growth, and fecundity in copepods and between bulk communities (i.e. phytoplankton, zooplankton) and in both lab and field settings (Mackas assumptions on rates of remineralization (McGillicuddy et al., and Bohrer, 1976; Baars and Oosterhuis, 1984; Dam et al., 1994; 1995a, b; Chai et al., 2003; Shulman et al., 2005). In order to Pasternak, 1994; Atkinson et al., 1996; Landry et al., 1997; Harris advance these model approaches, systematic measures of the et al., 2000). Jaffe et al. (1998) and Franks and Jaffe (2001, 2008) modeled quantities and their spatial and temporal scales of simultaneously imaged phytoplankton and zooplankton using a distributions are needed for initialization and model validation. fluorescence-imaging system and identified fluorescing zooplank­ Chlorophyll fluorescence and acoustics have been used to validate ton guts in situ. It follows, therefore, that a portion of in situ phytoplankton and zooplankton distributions, respectively; how­ fluorescence measurement would be attributable to zooplankton ever, there is presently no straightforward single measurement to gut fluorescence, with that contribution varying based on the level accomplish this over large domains. This study identifies a of trophic interaction and grazing. A systematic method of potential biological measurement that can be made on the space separating the fluorescence of living phytoplankton from zoo­ (kilometers) and time (days) scales relevant for model data plankton guts, however, has not been identified. The depth . As with formulating oceanographic models, the distribution of the 2002 fluorescence data from the high-CV data upstream conditions need to be considered when defining (Fig. 5) showed a pattern supportive of fluorescence being deepest Monterey Fig. 10. Depth point Bay over (see distributions Fig. Monterey 3) on of nine Canyon. fluorescence

successive August 18 August 17 August 16 August 15 August 14 August 13 August 12 August 11 August 10

Data Depth (m) -40 -30 -20 -10 are (RFU) 5 01 02 03 40 35 30 25 20 15 10 5 0 nights lacking greater from for August than the night 0.1 10 RFU to of 18, August where 2003. Distance (km) 14, CV The is 2006, greater dotted where line than the separates 0.70 fluorometer collected the two on by transects, the board REMUS the indicating AUV AUV had 0 1 2 3 along a the

poor Fluorescence (RFU) the furthest connection ‘‘a’’ and point ‘‘b’’ with offshore transect the and vehicle. lines the in attributed to gut contents of the zooplankton (Fig. 9). The References distribution of these data framed the nearshore autotrophic community from both the bottom and vertical fluorescence front Abrahams, V.A., Townsend, L.D., 1993. Bioluminescence in dinoflagellates: a test of 10 km offshore. Four percent of the total fluorescence along the the burglar alarm hypothesis. 74, 258–260. Alberte, R.S., 1993. Bioluminescence: the fascination, phenomena, and funda­ transect was found to be associated with the zooplankton CV mentals. Naval Research Reviews 45, 2–12. signal. Alldredge, A.L., Passow, U., Haddock, S.H.D., 1998. The characteristics and This finding has a number of significant implications. Fluores­ transparent exopolymer (TEP) content of marine snow formed from cence is used by the oceanographic community as a bulk measure thecate dinoflagellates. Journal of Plankton Research 20, 393–406. Atkinson, A., Shreeve, R.S., Pakhomov, E.A., Priddle, J., Blight, S.P., Ward, P., 1996. of phytoplankton biomass and subsequently used in estimates of Zooplankton response to a phytoplankton bloom near South Georgia, . If a significant fraction of this biomass was . Marine Ecology Progress Series 144, 195–210. actually in zooplankton guts and no longer viable for carbon Baars, M.A., Oosterhuis, S.S., 1984. Diurnal feeding rhythms in North copepods measured by gut fluorescence, digestive activity and grazing on fixation, there would be an overestimation of carbon productivity. labeled food. Netherlands Journal of Sea Research 8, 97–119. Additionally, these interactions are not distributed uniformly in Batchelder, H.P., Swift, E., 1989. Estimated near-surface mesoplanktonic biolumi­ the water column, leading to further complexity in application. nescence in the western North Atlantic during July 1986. and Oceanography 34, 113–128. Fig. 10 illustrates this further when the approach is applied to the Batchelder, H.P., Swift, E., van Keuren, J.R., 1990. Pattern of planktonic biolumines­ AOSN-II data set from Monterey Bay. As with the distribution of cence in the northern : seasonal and vertical distribution. Marine CV (Fig. 8B), it is clear that initially there was connectivity in Biology 104, 153–164. Batchelder, H.P., Swift, E., van Keuren, J.R., 1992. Diel patterns of planktonic grazing between the coast and the center of the bay. As the eddy bioluminescence in the northern Sargasso Sea. 113, 329–339. intensified, the coastal zooplankton separated from those in the Benoit-Bird, K.J., Au, W.W.L., 2003. Echo strength and density structure of Hawaiian center of the bay and developed distinct regions of trophic mesopelagic boundary community patches. Journal of the Acoustical Society of America 114, 1888–1897. transfer. By the last sampling night (into the relaxation period), Blackwell, S.M., 2002. A new platform for studying bioluminescence in the coastal there is some indication of the grazing connectivity returning. In a ocean. M.S. Thesis, California Polytechnic State University, California, USA, modeling context, delineating these fields has the potential to unpublished. advance coupled physical–biological models and refine the space Blackwell, S.M., Moline, M.A., Schaffner, A., Chang, G., 2007. Sub-kilometer length scales of physical and biological parameters in coastal waters. Continental and time scales of trophic interactions, carbon transfer, and rates Shelf Research 28, 215–226. of biogeochemical cycling. Although controlled experiments are Burkenroad, M.D., 1943. A possible function of bioluminescence. Journal of Marine clearly needed to fully validate the ability to discriminate Research 5, 161–164. Case, J.F., Widder, E.A., Bernsein, S., Ferer, K., Young, D., Latz, M.I., Geiger, M., Lapota, zooplankton gut fluorescence from viable phytoplankton biomass, D., 1993. Assessment of marine bioluminescence. Naval Research Reviews 45, results from this study suggest that the combined measurement 31–41. of bioluminescence and fluorescence may be used to delineate Chai, F., Jiang, M., Barber, R.T., Dugdale, R.C., Chao, Y., 2003. Interdecadal CViation of the transition zone chlorophyll front, a physical–biological model simulation regions of trophic interaction. between 1960 and 1990. Journal of Oceanography 59, 461–475. Clarke, G.L., Kelley, M.G., 1965. Measurements of diurnal changes in biolumines­ cence from the sea surface to 2,000 meters using a new photometric device. 4. Conclusions Limnology and Oceanography 10, 54–66. Clarke, G.L., Wertheim, G.K., 1956. Measurements of illumination at great depths and at night in the Atlantic Ocean by means of a new bathyphotometer. Deep- This study takes advantage of the general differences in Sea Research 3, 189–205. bioluminescence flash kinetics between dinoflagellates and Cochrane, N.A., Sameoto, D.D., Herman, A.W., 2000. Scotian Shelf euphausiid and silver hake population changes during 1984–1996 measured by multi­ zooplankton to measure the relative abundances of the two frequency acoustics. Journal of Marine Science 57, 122–132. groups within the same time space volume. Results demonstrate Dam, H.G., Peterson, W.T., Bellantoni, D.C., 1994. Seasonal feeding and fecundity of this as an approach for distinguishing these general classifications the calanoid copepod tonsa in Long Island Sound: is omnivory important to production? Hydrobiologia 292/293, 191–199. using a single sensor, which was validated in this study using Deutschman, D.H., Bradshaw, G.A., Childress, W.M., Daly, K.L., Grunbaum, D., fluorescence and exclusion experiments. Applied to large field Pascual, M., Schumaker, N., Wu, J., 1993. Mechanisms of patch formation. In: data sets, this approach has the potential to provide models with Levin, S., Powell, T., Steele, J. (Eds.), . Lecture Notes in Biomathematics, vol. 96. Springer, Berlin, pp. 184–208. distributions of these planktonic communities for initialization Ducklow, H.W., McCallister, S.L., 2005. The biogeochemistry of in and validation leading to mechanisms governing the patch the coastal oceans. In: Robinson, A.R., Brink, K.H. (Eds.), The Global Coastal distribution, coherence, and their biological interactions. The Ocean Multiscale Interdisciplinary Processes. The Sea, vol. 13. Harvard University Press, Cambridge, MA, pp. 269–315. organismal diversity and variability represented in the measure of Flagg, C.N., Smith, S.L., 1989. On the use of the acoustic doppler current profiler to bioluminescence at a given time and place prevents this approach measure zooplankton abundance. Deep-Sea Research 36, 455–474. from being quantitatively applied universally, but may be useful Franks, P.J.S., Jaffe, J.S., 2001. Microscale distributions of phytoplankton: initial on relatively short time and space scales. Despite these limita­ results from a two-dimensional imaging fluorometer, OSST. Marine Ecology Progress Series 220, 59–72. tions, the measure of BP may afford the oceanographic commu­ Franks, P.J.S., Jaffe, J.S., 2008. Microscale variability in the distributions of large nity a complimentary tool to observe and understand planktonic fluorescent particles observed in situ with a planar laser imaging fluorometer. communities in the ocean. Journal of Marine Systems 69, 254–270. Geistdoerfer, P., Cussatlegras, A.-S., 2001. nycthe´me´rales de la bioluminescence marine en Me´diterrane´e et dans l’Atlantique nord-est. Comptes Rendus de l’Acade´mie des Sciences 324, 1037–1044. Haddock, S.H.D., 1998. Bioluminescence in the deep-sea and open ocean: gelatinous zooplankton and marine snow. Ph.D. Thesis, University of California, Acknowledgements Santa Barbara, USA, unpublished. Haddock, S.H.D., Case, J.F., 1999. Bioluminescence spectra of shallow and deep-sea We thank a number of former graduate students (Jessica gelatinous zooplankton: ctenophores, medusae and siphonophores. Marine Biology 133, 571–582. Connolly, Ian Robbins, Michael Sauer) and undergraduate stu­ Haddock, S.H.D., Rivers, T.J., Robison, B.H., 2001. Can coelenterates make dents (Jeff Sevadjian, Carol Boland) for their help in sample ? Dietary requirement for in cnidarian biolumines­ collection and AUV deployments. We would also like to thank cence. Proceedings of the National Academy of Sciences of the United States of America 98, 11148–11151. Cyril Johnson for instrument calibration and design, and the crew Hall, S.J., Raffaelli, D.G., 1993. Food webs: theory and reality. Advances in Ecological of R/V Paragon (UCSC) and the AOSN-II team for their logistical Research 24, 187–239. support. This work was supported by the Office of Naval Research Harris, R., Wiebe, P., Lenz, J., Skjoldal, H-R., Huntley, M., 2000. ICES Zooplankton Methodology Manual. Academic Press, London, p. 684. (N00014-00-1-0570 and N00014-03-1-0341 to M. Moline). Herren, C.M., Alldredge, A.L., Case, J.F., 2003. Coastal bioluminescent marine snow: McGillicuddy, D.J., McCarthy, J.J., Robinson, A.R., 1995a. Coupled physical and fine structure of bioluminescence distribution. Continental Shelf Research 24, biological modeling of the in the North Atlantic: I. Model 413–442. formulation and one dimensional bloom processes. Deep-Sea Research I 42, Herren, C.M., Haddock, S.H.D., Johnson, C., Orrico, C.M., Moline, M.A., Case, J.F., 1313–1357. 2005. A multi-platform bathyphotometer for fine-scale, coastal biolumines­ McGillicuddy, D.J., Robinson, A.R., McCarthy, J.J., 1995b. Coupled physical and cence research. Limnology and Oceanography: Methods 3, 247–262. biological modeling of the spring bloom in the North Atlantic: II. Three Herring, P.J., 1987. Systematic distribution of bioluminescence in living organisms. dimensional bloom and post-bloom processes. Deep-Sea Research I 42, Journal of Bioluminescence and 1, 147–163. 1359–1398. Holliday, D.V., Pieper, R.E., 1980. Volume scattering strengths and zooplankton Moline, M.A., Heine, E., Case, J., Herren, C., Schofield, O., 2001. Spatial and temporal distributions at acoustic frequencies between 0.5 and 3 MHz. Journal of the CViability of bioluminescence potential in coastal regions. In: Case, J.F., Acoustical Society of America 67, 135–146. Herring, P.J., Haddock, S.H.D., Kricka, L.J., Stanley, P.E. (Eds.), Bioluminescence Holliday, D.V., Pieper, R.E., Kleppel, G.S., 1989. Determination of zooplankton size and Chemiluminescence 2000. World Scientific Publishing Company, Singa­ and distribution with multi-frequency acoustic technology. Journal of the pore, pp. 123–126. International Council for the Exploration of the Sea 41, 226–238. Moline, M.A., Blackwell, S.M., Von Alt, C., Allen, B., Austin, T., Case, J., Forrester, N., Holligan, P.M., Reiners, W.A., 1992. Predicting the responses of the coastal zone to Goldsborough, R., Purcell, M., Stokey, R., 2005. Remote environmental global change. Advances in Ecological Research 22, 211–255. monitoring units: an autonomous vehicle for characterizing coastal environ­ Jaffe, J.S., Franks, P.J.S., Leising, A.W., 1998. Simultaneous imaging of phytoplankton ments. Journal of Atmospheric and Oceanic Technology 22, 1797–1808. and zooplankton distributions. Oceanography 11, 1–5. Moore, C., 1994. Spectral absorption, , fluorescence meters—a new Jahnke, R.A., 2005. Transport processes and cycling in coastal window of opportunity for ocean scientists. Sea Technology 35, 10–16. sediments. In: Robinson, A.R., Brink, K.H. (Eds.), The Global Coastal Ocean Morin, J.G., 1983. Coastal bioluminescence: patterns and functions. Bulletin of Multiscale Interdisciplinary Processes. The Sea, vol. 13. Harvard University Marine Science 33, 787–817. Press, Cambridge, MA, pp. 163–191. Morin, J.G., Cohen, A.C., 1991. Bioluminescent displays, courtship, and Johnson, C.H., Knight, M., Trewavas, A., Kondo, T., 1998. A clockwork green: in ostracodes. In: Bauer, R., Martin, J. (Eds.), Crustacean Sexual Biology. circadian programs in photosynthetic organisms. In: Lumsden, P., Millar, A. Columbia University Press, New York, pp. 1–16. (Eds.), Biological Rhythms and Photoperiodism in . BIOS Scientific Morse, D., Milos, P.M., Roux, E., Hastings, J.W., 1989. Circadian regulation of the Publishers, Oxford, pp. 1–34. synthesis of substrate binding protein in the Gonyaulax bioluminescent Johnson, W.S., Allen, D.M. (Eds.), 2005. Zooplankton of the Atlantic and Gulf Coasts: system involves translational control. Proceedings of the National Academy A Guide to Their Identification and Ecology. Johns Hopkins Press, Baltimore, of Sciences of the United States of America 86, 172–176. p. 365. Pasternak, A.F., 1994. Gut fluorescence in herbivorous copepods: an attempt to Johnson, W.W., 1948. Sound as a tool in marine ecology, from data on biological justify the method. Hydrobiologia 292/293, 241–248. noises and the . Journal of Marine Research 7, 443–458. Petruncio, E.T., Rosenfeld, L.K., Paduan, J.D., 1998. Observations of the internal Kantha, L.H., Clayson, C.A. (Eds.), 2000. Small Scale Processes in Geophysical Flows. in Monterey Canyon. Journal of Physical Oceanography 28, 1873–1903. Academic Press, London, p. 888. Pieper, R.E., Holliday, D.V., Kleppel, S., 1990. Quantitative zooplankton distributions Kirkpatrick, G.J., Schofield, O., Millie, D.F., Moline, M.A., 2000. Optical discrimina­ form multifrequency acoustics. Journal of Plankton Research 12, 433–441. tion of a phytoplankton species in natural mixed populations. Limnology and Rees, J.-F., Wergifosse, B., Noiseti, O., Dubuissoni, M., Janssens, B., Thompson, E.M., Oceanography 45, 467–471. 1998. The origins of marine bioluminescence: turning defence Landry, M.R., Barber, R.T., Bidigare, R.R., Chai, F., Coale, K.H., Dam, H.G., Lewis, M.R., mechanisms into deep-sea communication tools. Journal of Experimental Lindley, S.T., McCarthey, J.J., Roman, M.R., Stoecker, D.K., Verity, P.G., White, J.R., Biology 201, 1211–1221. 1997. and grazing constraints on primary production in the Central Schofield, O., Bergmann, T., Bissett, W.P., Moline, M.A., Orrico, C., 2004. Inverting Equatorial Pacific: an EqPac synthesis. Limnology and Oceanography 42, inherent optical signatures in the nearshore coastal waters at the Long Term 405–418. Ecosystem Observatory. Journal of Geophysical Research 109, C12S04. Lapota, D., 1998. Long term and seasonal changes in dinoflagellate biolumines­ Seliger, H.H., Fastie, W.G., McElroy, W.D., 1969. Towable photometer for rapid area cence in the Southern California Bight. Ph.D. Thesis, University of California, mapping of concentrations of bioluminescent marine dinoflagellates. Limnol­ Santa Barbara, USA, unpublished. ogy and Oceanography 14, 806–813. Lapota, D., Losee, J.R., 1984. Observations of bioluminescence in marine plankton Shulman, I., Haddock, S.H.D., McGillicuddy, D.J., Paduan, J.D., Bissett, W.P., 2003. from the Sea of Cortez. Journal of Experimental Marine Biology and Ecology 77, Numerical modeling of bioluminescence distributions in the coastal ocean. 209–240. Journal of Atmospheric and Oceanic Technology 20, 1060–1068. Lapota, D., Galt, C., Losee, J., Huddell, H.D., Orzech, J.K., Nealson, K.H., 1988. Shulman, I., McGillicuddy Jr., D.J., Moline, M.A., Haddock, S.H.D., Kindle, J.C., Observations and measurements of planktonic bioluminescence in and around Nechaev, D., Phelps, M.W., 2005. Bioluminescence intensity modeling and a milky sea. Journal of Experimental Marine Biology and Ecology 119, sampling strategy optimization. Journal of Atmospheric and Oceanic Technol­ 55–81. ogy 22, 1267–1281. Lapota, D., Geiger, M.L., Stiffey, A.V., Rosenberger, D.E., Young, D.K., 1989. Soli, G., 1966. Bioluminescent cycle of photosynthetic dinoflagellates. Limnology Correlation of planktonic bioluminescence with other oceanographic para­ and Oceanography 11, 355–363. meters from a Norwegian fjord. Marine Ecology Progress Series 55, 217–228. Swift, E., Sullivan, J.M., Batchelder, H.P., Van Keuren, J., Vaillancourt, R.D., Bidgare, Lapota, D., Rosenberger, D.E., Lieberman, S.H., 1992. Planktonic bioluminescence in R.R., 1995. Bioluminescent organisms and bioluminescence measurements in the pack ice and the marginal ice zone of the . Marine Biology 112, the North Atlantic Ocean near latitude 59.5N, longitude 21W. Journal of 665–675. Geophysical Research 100, 6527–6547. Latz, M.I., Frank, T.M., Case, J.F., 1988. Spectral composition of bioluminescence of Thomas, C.R., 1997. Identifying marine Phytoplankton. Academic Press, San Diego, epipelagic organisms from the Sargasso Sea. Marine Biology 98, 441–446. CA, p. 858. Levin, S.A., 1992. The problem of pattern and scale in ecology. Ecology 73, Widder, E.A., Case, J.F., 1981. Two flash forms in the bioluminescent dinoflagellate, 1943–1967. . Journal of Comparative Acta 143, 43–52. Lieberman, S.H., Lapota, D., Losee, J., Zirino, A., 1987. Planktonic bioluminescence in Widder, E.A., Case, J.F., Bernstein, S.A., MacIntyre, S., Lowenstine, M.R., Bowlby, the surface waters of the Gulf of California. 4, 25–46. M.R., Cook, D.P., 1993. A new large volume bioluminescence bathyphotometer Lorenzen, C.J., 1966. A method of the continuous measurement of in vivo with defined turbulence excitation. Deep-Sea Research 40, 607–627. Chlorophyll concentration. Deep-Sea Research 13, 223–227. Widder, E.A., Johnsen, S., Bernstein, S.A., Case, J.F., Neilson, D.J., 1999. Thin layers of Losee, J., Richter, K., Lieberman, S., Lapota, D., 1989. Bioluminescence: spatial bioluminescent copepods found at density discontinuities in the water statistics in the North Atlantic. Deep-Sea Research 36, 783–801. column. Marine Biology 134, 429–437. Mackas, D., Bohrer, R., 1976. Fluorescence analysis of zooplankton gut contents and Yentsch, C.S., Menzel, D.W., 1963. A method for the determination of phytoplank­ an investigation of diel feeding patterns. Journal of Experimental Marine ton chlorophyll and phaeophytin by fluorescence. Deep-Sea Research 10, Biology and Ecology 25, 77–85. 221–231.